CN117134946A - Access request current limiting methods, devices, equipment and storage media - Google Patents

Abstract

The application relates to the technical field of communication and provides a method, a device, equipment and a storage medium for limiting access requests. The method comprises the following steps: acquiring a tag identifier of an access request of a back-end service; the tag identifier is used for indicating an initial current limit value corresponding to the access request and a first target server accessed by the access request; determining target requests of which the request quantity exceeds an initial current limit value corresponding to the tag identification; inquiring the resource use information of the first target server, and limiting the current of the target request according to the resource use information; the resource usage information includes processor resource usage, disk usage, and memory usage. According to the access request current limiting method provided by the embodiment of the application, the label identification is introduced, the initial current limiting value based on the label identification is used for current limiting, and the dynamic current limiting of the access request exceeding the initial current limiting value based on the resource use condition is combined, so that the utilization rate of the resource can be improved.

Description

Access request current limiting methods, devices, equipment and storage media

Technical field

The present application relates to the field of communication technology, and specifically relates to an access request current limiting method, device, equipment and storage medium.

Background technique

As the application of microservices becomes more and more widespread, due to the interdependence between the business process interfaces of each microservice, complex dependencies are formed where many platforms call the same platform. If there is a problem with a certain request of a certain system, This will lead to abnormal calls and may cause the entire service to be paralyzed. To this end, it is necessary to implement flexible isolation and current limiting processing for businesses under the microservice framework. At the same time, although current limiting can make the system very stable, it will reduce system throughput, reduce resource utilization, and increase server costs to a certain extent. For example, when the business is busy, the query rate per second (QPS) is 1000, and in another idle period, the QPS is 3000. In order to allow the system to stably provide services, the current limiting setting will set the QPS to 1000, which makes the A time period with a QPS of 3,000 can only provide service capabilities with a QPS of 1,000.

Existing current limiting methods are all static current limiting methods, such as leaky bucket, token bucket, counter method, etc., which can limit the number of requests per unit time and ensure the normal operation of the system. However, these methods all unify requests. The current limit is more granular and only limits the system interface. There is no isolation and differentiation of the requested service or system. It is easy to cause an exception in a certain request to affect the requests of other systems. Based on this, some existing current limiting methods configure the priority of requests, respond to high-priority service requests first, and then respond to low-priority service requests. When the amount of concurrency is not high, this method reserves too many resources for high-priority requests. Even if the service is idle, low-priority requests will still have the problem of unresponsive requests, which will greatly reduce resource utilization.

Contents of the invention

Embodiments of the present application provide an access request current limiting method, device, equipment, and storage medium to solve the technical problem of low resource utilization in existing access request current limiting methods.

In the first aspect, embodiments of this application provide a method for limiting access request current, including:

Obtain the label identification of the access request of the back-end service; the label identification is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request;

Determine the target request whose number of requests for the access request exceeds the initial current limit value corresponding to the tag identification;

Query the resource usage information of the first target server, and limit the flow of the target request according to the resource usage information; the resource usage information includes processor resource usage, disk usage, and memory usage.

In one embodiment, the tag identification is also used to indicate the resource type required by the access request, and the resource type includes processor resources, memory resources and disk resources; the resource usage information is used to determine the resource type. Target requests are flow-limited, including:

Determine the target resource type required by the target request according to the tag identification;

Determine remaining resources of the first target server according to the resource usage information;

Based on the target resource type and the remaining resources, determine whether the first target server has available resources corresponding to the target request;

If it does not exist, obtain the target resource type and query whether there is a second target server in the back-end service; the second target server is the one that exists in the back-end service in addition to the first target server. The server of available resources corresponding to the target request;

If there is a second target server in the backend service, the target request is flow-limited according to the available resources of the second target server.

In one embodiment, the tag identifier is also used to indicate the timestamp information of the access request; and limiting the target request according to the available resources of the second target server includes:

Obtain the target timestamp of the target request under the tag identification;

Determine the window size based on the available resources of the second target server;

Sort the target requests according to the order corresponding to the target timestamp;

According to the sorting order of the target requests, a sliding window algorithm is used to respond to the target requests based on the window size to limit the flow of the target requests.

In one embodiment, query whether the second target server in the backend service has available resources corresponding to the target request according to the following formula:

R=αC+βM+δD;

Among them, C is the processor usage, M is the memory usage, D is the disk usage, R is the resource abundance, and E _i is the i-th server in the back-end service except the first target server. The remaining resource indicator value, busness is the resource type preference factor corresponding to the target resource type, and fail _count is the number of failed requests within the historical preset time period; ε, α, β and δ are all preset weight parameters.

In one embodiment, determining that the number of request requests for the access request exceeds the target request corresponding to the initial current limit value corresponding to the tag identification includes:

Classify and summarize the access requests based on the label identification, and classify and summarize the access requests with the same label identification under the same label classification;

According to the classification summary results, determine the number of access requests under the target label classification; the target label classification is any one of the label classifications corresponding to the label identification;

If the number of requests exceeds the initial current limit value corresponding to the tag identifier under the target tag category, obtain the target requests under the target tag category that exceed the initial current limit value.

In one embodiment, before obtaining the tag identifier of the access request for the backend service, the method further includes:

Obtain the request information of the access request of the back-end service; the request information includes the first target server accessed by the access request, the resource type required by the access request, and the timestamp information of the access request;

Generate a label identification of the access request according to the request information.

In one embodiment, the access request current limiting method further includes:

Obtain resource usage information of historical access requests;

The initial current limit value corresponding to the tag identifier of the historical access request is determined according to the resource usage information.

In the second aspect, embodiments of the present application provide an access request current limiting device, including:

A request interception module, used to obtain the tag identification of the access request of the back-end service; the tag identification is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request;

A first current limiting module, configured to determine target requests whose number of access requests exceeds the initial current limiting value corresponding to the tag identification;

The second current limiting module is used to query the resource usage information of the first target server and limit the current flow of the target request according to the resource usage information; the resource usage information includes processor resource usage and disk usage. and memory usage.

In a third aspect, embodiments of the present application provide an electronic device, including a processor and a memory storing a computer program. When the processor executes the program, the steps of the access request current limiting method described in the first aspect are implemented.

In the fourth aspect, embodiments of the present application provide a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the access request current limiting method described in the first aspect is implemented. step.

In a fifth aspect, embodiments of the present application provide a computer program product, including a computer program that, when executed by a processor, implements the steps of the access request current limiting method described in the first aspect.

The access request current limiting method, device, equipment and storage medium provided by the embodiments of the present application obtain the tag identifier used to indicate the initial current limit value and the server requesting access, and determine that the number of access requests exceeds the initial number corresponding to the tag identifier. Target request with current limit value; query resource usage information such as processor resource usage, disk usage, and memory usage of the first target server, limit the target request based on the resource usage information, and perform the remaining availability resources of the server. Flexible utilization enables dynamic flow limitation of requests and improves resource utilization.

Description of the drawings

In order to explain the technical solutions in this application or the prior art more clearly, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are of the present invention. For some embodiments of the application, those of ordinary skill in the art can also obtain other drawings based on these drawings without exerting creative efforts.

Figure 1 is one of the flow diagrams of the access request current limiting method provided by the embodiment of the present application;

Figure 2 is a schematic diagram of the tag identification control bit provided by the embodiment of the present application;

Figure 3 is a schematic diagram of the tag identification of the access request provided by the embodiment of the present application;

Figures 4 to 5 are schematic diagrams of label classification of label identification provided by embodiments of the present application;

Figure 6 is the second schematic flowchart of the access request current limiting method provided by the embodiment of the present application;

Figure 7 is the third schematic flowchart of the access request current limiting method provided by the embodiment of the present application;

Figure 8 is a schematic structural diagram of an access request current limiting device provided by an embodiment of the present application;

Figure 9 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.

Detailed ways

In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are part of this application. Examples, not all examples. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this application.

Figure 1 is a schematic flowchart of an access request current limiting method provided by an embodiment of the present application. Referring to Figure 1, the access request current limiting method provided by the embodiment of the present application may include:

Step 100: Obtain the tag identifier of the access request of the back-end service; the tag identifier is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request;

Obtain the tag ID (serviceID) of the access request of the backend service. The tag ID is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request. The tag ID is also used to indicate the requested service, and The business system to be called by the service, and the requested service may correspond to the first target server. In the cluster service under the microservice framework, one business system can call another one or more business systems through access requests, and the same business system can also be called by one or more other business systems, thus forming a complex situation in the cluster. Calling dependencies, different services need to call different business systems, thus requiring calls to different interfaces.

Furthermore, the label identification may be generated by mapping the request information of the access request based on preconfigured label identification mapping rules when the access request is initiated. Specifically, in the label identification mapping rules, different business systems, different servers, different resource types, and different interfaces correspond to different identifiers. Based on the preset label identification mapping rules, the business to be called according to the access request The system, and/or the requested service, and/or the required resource type, and/or the interface called, generates the label identification of the access request.

It should be noted that the label identification is generated by adding the label control bit as shown in Figure 2 to the header of the message body of the access request. The label control bits of the access request are the same, indicating that the requests belong to the same label. Each system can generate the same or multiple access requests with different labels. The same business system generates access requests containing multiple different labels. The label identification is shown in Figure 3. In Figure 3, the access request to be called The business system includes system A and system B. Therefore, the tag identifier in the generated access request includes the identifier serviceIDa corresponding to system A and the identifier serviceIDb corresponding to system B.

Step 200: Determine target requests whose number of access requests exceeds the initial current limit value corresponding to the tag identification;

Determine the target request whose number of access requests exceeds the initial current limit value corresponding to the tag. It can be understood that when the number of access requests does not exceed the initial current limit value, you can directly respond to the access request. When the number of access requests When the number of requests exceeds the initial current limit value, the system will respond to requests that do not exceed the initial current limit value, and for requests that exceed the initial current limit value, current limit processing is required. For target requests that exceed the initial current limit value, the access request can be selected based on the time sequence of the requests and/or the priority of the requests, provided that the initial current limit value is not exceeded, and the request time is first among the access requests, and/or , the request with the highest priority will be responded to, and the request time will be later and/or the priority will be lower, and the current will be limited as the target request that exceeds the initial current limit value.

Furthermore, the initial current limit values corresponding to different label identifiers can be the same or different. The required resource types and required resource amounts for access requests under different label identifiers can also be different. In cluster services, according to Different label identifiers require resource types and resource amounts, and different initial current limit values are assigned to different label identifiers. When access requests under the same label identifier exceed its initial current limit value, target requests that exceed the initial current limit value will be processed. Perform current limiting.

Step 300: Query the resource usage information of the first target server, and limit the flow of the target request according to the resource usage information; the resource usage information includes processor resource usage, disk usage, and memory usage.

When limiting the target request, specifically, the resource usage information of the first target server is queried, and the target request is limited based on the resource usage information. Specifically, if the first target server has resources available for the target request, use the resources of the first target server to respond to the target request; if the first target server does not have resources available for the target request, select resources from the back-end service. The server with the most remaining responds to the target request and implements dynamic current limiting of the target request.

Furthermore, based on the request information indicated by the tag identifier, the tag identifier can be used to limit the flow of access requests in different dimensions based on any one or more items in the request information, thereby limiting the flow of access requests to the back-end service. Granularity can be flexibly controlled to achieve dynamic current limiting.

In this embodiment, by obtaining the tag identifier indicating the initial current limit value and the server requesting access, and determining the target request whose number of access requests exceeds the initial current limit value corresponding to the tag identifier; querying the first target server Resource usage information such as processor resource usage, disk usage, and memory usage can be used to limit the target request based on the resource usage information. The remaining availability resources of the server can be flexibly utilized to achieve dynamic flow limitation of requests and improve resources. Utilization.

Preferably, in step 100, before obtaining the tag identification of the access request of the back-end service, it is first necessary to obtain the access request of the back-end service and generate the tag identification corresponding to the access request. Specifically, before obtaining the label identification of the access request of the backend service, you may also include:

Step 001, obtain the request information of the access request of the back-end service; the request information includes the first target server accessed by the access request, the resource type required by the access request, and the timestamp information of the access request;

Step 002: Generate a tag identification of the access request according to the request information.

Obtain the access request of the back-end service and its request information. The request information includes the first target server accessed by the access request, the resource type required by the request and the timestamp information of the request. Based on the request information of the access request, generate the access request Label identification. The label identification can be generated by mapping the request information based on preset rules, or it can be obtained by matching the request information of the access request with the corresponding relationship table based on the correspondence table between the request information and the label identification set in advance. There is no specific limitation here.

Preferably, the initial current limiting value corresponding to the tag identification is determined based on the resource usage of the access request under each tag identification. The access request current limiting method provided by the embodiment of the present application may also include:

Step 01: Obtain the resource usage information of historical access requests;

Step 02: Determine the initial current limit value corresponding to the tag identifier of the historical access request based on the resource usage information.

Obtain the resource usage information of historical access requests. Different tag identifiers require different resource types. Tag identifiers that require the same resource type can also require different resource amounts for access requests. Based on the resource usage information of historical access requests, determine the resource usage of access requests under each label identification, and then allocate available resources to each label identification based on the resource situation of the cluster service, and set the initial current limit value corresponding to each label identification.

In one embodiment, different tag identifiers correspond to different initial current limit values. The initial current limit value corresponding to the tag identifier can be determined based on a pre-designed tag resource table. In the pre-designed tag resource table, each tag The identifier corresponds to an initial current limit value, which is determined based on the resource usage of access requests under each tag identifier in historical access requests. Specifically, each tag identifier sets an initial preset QPS according to its usage of resources. _Ti is the static initial value of the tag identifier i, that is, the initial current limit value of the tag identifier i. The actual QPS of the backend service is A _i (A _i ≥ T _i ). The initial preset QPS initial value can be used for stress test evaluation when setting, or a model can be established for parameter tuning through genetic algorithms.

Preferably, the obtained target requests that exceed the initial current limit value corresponding to the tag identifier are based on the label classification of the tag identifier, and obtain the target requests that exceed the initial current limit value under each label classification. In step 200, determining the target request whose number of access requests exceeds the initial current limit value corresponding to the tag identification may also include:

Step 201: Classify and summarize the access requests based on the label identification, and classify and summarize the access requests with the same label identification under the same label classification;

Step 202: Determine the number of access requests under the target label classification according to the classification and summary results; the target label classification is any one of the label classifications corresponding to the label identification;

Step 203: If the number of requests exceeds the initial current limit value corresponding to the tag identifier under the target tag category, obtain the target requests under the target tag category that exceed the initial current limit value.

When determining the target request whose number of access requests exceeds the initial current limit value corresponding to the tag ID, first classify and summarize the access requests based on the tag ID, and classify and summarize the access requests with the same tag ID under the same tag category. The tag classification is the same as the tag ID. Tag identification has a corresponding relationship. Under different current limiting control dimensions, the label classification corresponding to the same label identification may be different. However, under a specific current limiting control dimension, the label classification and label identification have a one-to-one correspondence.

Based on the classification and summary results of the access requests, the number of access requests under the target label classification is determined, and the target label classification is any one of the label classifications corresponding to the label identification. That is, based on the classification and summary results of access requests, determine the number of requests under each label classification, that is, the number of access requests with the same label identification. If the number of access requests under the same label identification exceeds the number corresponding to the label identification, Initial current limit value, then obtain the target requests under the target label category that exceed the initial current limit value.

In one embodiment, refer to Figure 4, which is a schematic diagram of the label classification method provided by the embodiment of the present application. Based on Figure 4, each business system in the microservice cluster serves as a calling system, initiates an access request to the server, and calls the corresponding service , the access request calling the system is forwarded to the server through the load balancer F5 for response. Among them, F5 acts as a load balancer to distribute traffic to multiple servers to improve system availability. The access requests of each calling system are classified and summarized according to the label identification of the access requests. The label classification method of the access requests of each calling system is shown in Figure 5.

Based on the label classification methods shown in Figures 4 and 5, if the label identification of the request interface of system A contains {Q ₁ ,…,A,…,Q _n }, the label identification of the request interface of system B is {B ₁ , …A,…,B _n }, the request interfaces of system A and system B contain the same label identification set {A}, but also contain different label sets. Based on the label classification method in the access request message body, the control granularity of current limiting is more precise and flexible.

Further, as shown in Figure 4, the label identification set of the n systems called is A{A ₁ ,A ₂ ,...,A _n }, and each label corresponds to a subset U{u _i of the set A, ..., u _j }. For example, the label identification of calling system A and the label of calling system B are both serviceIDab, and the label of calling system C is serviceIDc. The classification method based on the call system can make the granularity of current limiting more precise and flexible at the system level.

Preferably, the tag identification of the access request is also used to indicate the resource type required for the access request. The resource type includes processor resources, memory resources and disk resources. It can be seen that different access requests correspond to different services and the required resources The types are different. Therefore, access requests under different label identifications have different resource type preferences. Further, when limiting the target request under the same tag identification, the target request is dynamically limited based on the sliding window algorithm. In step 300, the target request is limited according to the resource usage information of the first target server. Streams, specifically including:

Step 301: Determine the target resource type required by the target request according to the tag identification;

Step 302: Determine remaining resources of the first target server based on the resource usage information;

Step 303: Based on the target resource type and the remaining resources, determine whether the first target server has available resources corresponding to the target request;

Step 304: If it does not exist, obtain the second target server in the back-end service based on the target resource type; the second target server is one of the back-end services except the first target server. In addition, there is a server with available resources corresponding to the target request;

Step 305: If there is a second target server in the backend service, limit the flow of the target request according to the available resources of the second target server.

Determine the target resource type required by the target request according to the tag identification of the target request, determine the remaining resources of the first target server based on the resource usage information of the first target server, and determine the target resource type required by the target request and the first target server's remaining resources. For remaining resources, determine whether the first target server has resources available for the target request. If there are resources available for the target request, perform a current limiting response to the target request based on the first target server; if the first target server does not have resources available for the target request, or the first If the remaining resources of the target server are less than the resources required by the target request, the backend service is queried to see if there is a second target server. If a second target server exists, a current-limiting response to the target request is performed based on the available resources of the second target server.

Further, the second target server is a server in the back-end service other than the first target server that has available resources corresponding to the target request. Optionally, the second target server is a server in the back-end service that has available resources corresponding to the target request. The server with the most resources.

Preferably, the tag identification of the access request is also used to indicate the timestamp information of the access request. In step 305, limiting the target request according to the available resources of the second target server may also include:

Step 315: Obtain the target timestamp of the target request under the tag identification;

Step 325, determine the window size according to the available resources of the second target server;

Step 335: Sort the target requests according to the order corresponding to the target timestamp;

Step 345: According to the sorting order of the target requests, a sliding window algorithm is used to respond to the target requests based on the window size to limit the flow of the target requests.

When performing a current-limiting response to a target request based on the available resources of the second target server, first obtain the timestamp information of the target request, that is, the target timestamp, which is generated when the target request enters. The window size is determined according to the available resources of the second target server, and the target requests are sorted according to the time sequence corresponding to the timestamp of the target request. According to the sorting order, a sliding window algorithm is used to respond to the target request based on the window size to achieve the goal. Requested current limit.

When performing a current-limiting response to a target request, a sliding window algorithm is used to respond to a request within the window each time. The window size is determined based on the available resources of the second target server to ensure that the resources used to respond to the target request do not exceed the resources of the second target server. load.

Further, when determining whether there is a second target server in the back-end service, the remaining resource indicators of each server in the back-end service except the first target server are calculated according to the following formula 1-2. According to the calculated The indicator value is used to determine whether there is a second target server that meets the conditions, where Formula 1-2 is as follows:

R＝αC+βM+δD; (2)

Among them, C is the processor usage, M is the memory usage, D is the disk usage, R is the resource abundance, and E _i is the i-th server in the back-end service except the first target server. The remaining resource indicator value, busness is the resource type preference factor corresponding to the target resource type, and fail _count is the number of failed requests within the historical preset time period; ε, α, β and δ are all preset weight parameters.

When the target request is an access request under a certain tag ID, since each tag ID is assigned an initial current limit value, different types of resources are correspondingly allocated to each tag ID. In this case, Formula 1- In 2, C is the processor usage under the label, M is the memory usage under the label, D is the disk usage under the label, and busness is the required access request under the label. The resource type preference factor corresponding to the resource type.

In order to prevent the cluster service from failing to process requests due to damage to a certain node, but resources are still reported, the remaining situation indicator E introduces the number of failed requests for a preset time period in history. The greater the number of failed requests, the higher the probability that the server will be called later. The lower. Requests identified by different tags require different resource preferences. For example, some access requests consume CPU processor resources, and some access requests consume memory resources. The resource remaining indicator E introduces the request resource preference factor busness for different tag identifiers. Among the servers in the back-end service except the first target server, the overall factor of the remaining resources of the i-th server is E _i . The server with the largest E _i value is selected as the second target server and performs a current-limiting response to the target request. .

In one embodiment, dynamic current limiting of access requests is implemented based on a preset dynamic current limiting controller. The dynamic current limiting controller includes an interceptor, a current limiting controller and a dynamic adjustment module. Based on the dynamic current limiting controller The access request current limiting process is shown in Figure 6. As shown in Figure 6, the dynamic current limiting controller also stores resource information of the backend service. This resource information can be actively reported by each server in the backend service, or can be actively obtained by the dynamic current limiting controller. This is not specifically limited.

Based on the dynamic current limiting process shown in Figure 6, access requests from business systems such as calling system A, calling system B, and calling system C first enter the interceptor through load balancer F5, and then the request is forwarded to the current limiting controller. The interceptor is used to receive instructions from the current limiting controller, intercept access requests and forward them to the current limiting controller. The current limit controller controls the flow of the initial current limit value for each request identified by the tag. Requests that exceed the initial current limit value enter the dynamic adjustment module for dynamic current limit.

Further, the dynamic current limiting module is provided with a resource monitoring table. The resource monitoring table is generated based on the resource information of the back-end service. Based on the resource monitoring table, the resource remaining status of the nodes in the cluster is mapped to the resource remaining status indicator. Obtain the resource remaining status indicator value of each node server. If the requested server has available remaining resources, the requested server resources will be used to perform a current limiting response. If the requested server does not have available resources, the resources of the server with the most remaining resources will be used to perform a dynamic current limiting response.

When responding to dynamic current limiting, the cluster records the timestamp of each tag identifying serviceID, and uses a sliding window algorithm for current limiting control. Set the unique identifier in the cluster as a set sequence collection containing the label keyword key. By setting the key survival time of the cluster and based on the sliding window algorithm, the corresponding value of the requested key is set to the millisecond timestamp of the request entry time plus the count value of the auto-incremented key. For requests that exceed the initial current limit value QPS, they are uniformly sent to the dynamic adjustment module for dynamic current limit control. For static indicators: once set, they cannot be changed. For example, when the serviceID is 001, the QPS when the interface is called is 3000. However, considering the resource usage when serviceID is 002, the actual initial current limit value set is much less than 3000, so that resources are reserved for requests with serviceID 002. However, when the request with serviceID 001 exceeds its set initial current limit value and the number of requests with serviceID 002 is small, the resources reserved for requests with serviceID 002 will be wasted, resulting in reduced resource utilization. In this embodiment, based on the dynamic adjustment module, requests that exceed the initial current limit value are transferred to the dynamic adjustment module for secondary judgment, which can improve resource utilization.

Specifically, the dynamic adjustment module is equipped with a resource monitoring table. The remaining resources owned by the nodes in each cluster are mapped to the resource remaining indicators. Through the designed resource remaining indicators, it can be judged whether the server requested for access has any resources that can be used directly. resources, or select the server with the most available resources to increase utilization of available resources.

Referring to Figure 7, Figure 7 is a flow chart of the access request current limiting method applied to the SMS platform. In view of the scenario where multiple system calls exist in the cluster under the microservice framework, such as the SMS platform, the designed label classification method and based on The label's dynamic current limiting controller, combined with the sliding window algorithm, achieves more flexible and efficient real-time current limiting. In Figure 7, the label control bit is introduced, and the message body of the system or access request is defined according to the requirements, and the label-based granular control method is used to define labels for the message body of the system or access request. According to the usage of the tag, for the tag current limit table in the current limit controller, set the initial value of the initial current limit value QPS according to the tag. When setting the initial current limit value QPS, pressure testing and evaluation can be performed, and a model can also be established for parameter tuning through genetic algorithms.

Further, the message body of the access request carries the label and reaches the interceptor, which intercepts the access request and forwards it to the current limiting controller. The message body of the access request reaches the current limiting controller, and the current limiting controller limits the current flow of the message body of the access request according to the tag resource table. Current limiting uses the timestamp under each label recorded by the cluster, and uses a sliding window algorithm for current limiting control. If the actual number of requests for the same label is within the set initial current limiting value, the normal access interface will respond to the request directly. Otherwise, the message body of the access request that exceeds the initial current limit value will be forwarded to the dynamic adjustment module for processing using label-based dynamic current limit. Specifically, the resource monitoring table is queried to determine whether the server requested to be called has available remaining resources. If it exists, access is normal. If not, the resource monitoring table is continued to be queried to determine whether the server with the largest remaining resources has available remaining resources. Execute the request, and if it exists, access it normally. If there are still no remaining resources available, downgrade the access request and lower the priority of the request to wait for execution.

In this embodiment, by introducing the concept of tags and based on the characteristics of large system calls and large requests in cluster services, a corresponding tag classification method is designed to make the current limit more accurate and flexible. At the same time, a tag-based dynamic current limiting controller is used to set the initial current limiting value of the same type of tags through genetic algorithms or pressure tests. After the traffic enters, the sliding window algorithm is used to dynamically limit the flow. For the same The traffic whose class tag exceeds the initial current limit value will be judged twice through the dynamic adjustment module. The dynamic adjustment module is equipped with a resource monitoring table to monitor the remaining resource indicators. It can be based on the remaining resources of the requested server and other servers. Dynamically respond to traffic exceeding the initial throttling value, thereby improving the utilization of remaining resources.

Furthermore, the access request current limiting method provided by the embodiment of this application adopts a tag-based dynamic current limiting controller, which not only avoids the problem of low resource utilization caused by the static current limiting method, but also avoids the request caused by completely dynamic configuration. The problem of poor user experience caused by slow response speed improves resource utilization while ensuring user experience.

The access request current limiting device provided by the embodiment of the present application is described below. The access request current limiting device described below and the access request current limiting method described above may be mutually referenced.

Referring to Figure 8, the access request current limiting device provided by the embodiment of the present application includes:

The request interception module 10 is used to obtain the tag identification of the access request of the back-end service; the tag identification is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request;

The first current limiting module 20 is used to determine the target request whose number of requests for the access request exceeds the initial current limiting value corresponding to the tag identification;

The second current limiting module 30 is used to query the resource usage information of the first target server, and limit the current flow of the target request according to the resource usage information; the resource usage information includes processor resource usage, disk usage rate and memory usage.

In one embodiment, the tag identification is also used to indicate the resource type required for the access request, and the resource type includes processor resources, memory resources and disk resources; the second current limiting module 30 also uses At:

Determine the target resource type required by the target request according to the tag identification;

Determine remaining resources of the first target server according to the resource usage information;

Based on the target resource type and the remaining resources, determine whether the first target server has available resources corresponding to the target request;

If it does not exist, obtain the target resource type and query whether there is a second target server in the back-end service; the second target server is the one that exists in the back-end service in addition to the first target server. The server of available resources corresponding to the target request;

If there is a second target server in the backend service, the target request is flow-limited according to the available resources of the second target server.

In one embodiment, the tag identification is also used to indicate the timestamp information of the access request; the second current limiting module 30 is also used to:

Obtain the target timestamp of the target request under the tag identification;

Determine the window size based on the available resources of the second target server;

Sort the target requests according to the order corresponding to the target timestamp;

According to the sorting order of the target requests, a sliding window algorithm is used to respond to the target requests based on the window size to limit the flow of the target requests.

In one embodiment, query whether the second target server in the backend service has available resources corresponding to the target request according to the following formula:

R=αC+βM+δD;

Among them, C is the processor usage, M is the memory usage, D is the disk usage, R is the resource abundance, and E _i is the i-th server in the back-end service except the first target server. The remaining resource indicator value, busness is the resource type preference factor corresponding to the target resource type, and fail _count is the number of failed requests within the historical preset time period; ε, α, β and δ are all preset weight parameters.

In one embodiment, the first current limiting module 20 is also used to:

Classify and summarize the access requests based on the label identification, and classify and summarize the access requests with the same label identification under the same label classification;

According to the classification summary results, determine the number of access requests under the target label classification; the target label classification is any one of the label classifications corresponding to the label identification;

If the number of requests exceeds the initial current limit value corresponding to the tag identifier under the target tag category, obtain the target requests under the target tag category that exceed the initial current limit value.

In one embodiment, the access request current limiting device further includes a label generation module for:

Obtain the request information of the access request of the back-end service; the request information includes the first target server accessed by the access request, the resource type required by the access request, and the timestamp information of the access request;

Generate a label identification of the access request according to the request information.

In one embodiment, the access request current limiting device further includes an initial value generation module for:

Obtain resource usage information of historical access requests;

The initial current limit value corresponding to the tag identifier of the historical access request is determined according to the resource usage information.

Figure 9 illustrates a schematic diagram of the physical structure of an electronic device. As shown in Figure 9, the electronic device may include: a processor (processor) 910, a communication interface (Communication Interface) 920, a memory (memory) 930 and a communication bus 940. Among them, the processor 910, the communication interface 920, and the memory 930 complete communication with each other through the communication bus 940. The processor 910 can call the computer program in the memory 930 to perform the steps of the access request current limiting method, for example, including:

Obtain the label identification of the access request of the back-end service; the label identification is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request;

Determine the target request whose number of requests for the access request exceeds the initial current limit value corresponding to the tag identification;

Query the resource usage information of the first target server, and limit the flow of the target request according to the resource usage information; the resource usage information includes processor resource usage, disk usage, and memory usage.

In addition, the above-mentioned logical instructions in the memory 930 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the existing technology or the part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of this application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program code. .

On the other hand, embodiments of the present application also provide a computer program product. The computer program product includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. The computer program is executed by a processor. When , the computer can perform the steps of the access request current limiting method provided by the above embodiments, including, for example:

Obtain the label identification of the access request of the back-end service; the label identification is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request;

Determine the target request whose number of requests for the access request exceeds the initial current limit value corresponding to the tag identification;

Query the resource usage information of the first target server, and limit the flow of the target request according to the resource usage information; the resource usage information includes processor resource usage, disk usage, and memory usage.

On the other hand, embodiments of the present application also provide a processor-readable storage medium. The processor-readable storage medium stores a computer program. The computer program is used to cause the processor to execute the access request provided by the above embodiments. The steps of the current limiting method include, for example:

Obtain the label identification of the access request of the back-end service; the label identification is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request;

Determine the target request whose number of requests for the access request exceeds the initial current limit value corresponding to the tag identification;

Query the resource usage information of the first target server, and limit the flow of the target request according to the resource usage information; the resource usage information includes processor resource usage, disk usage, and memory usage.

The processor-readable storage medium may be any available media or data storage device that the processor can access, including but not limited to magnetic storage (such as floppy disks, hard disks, tapes, magneto-optical disks (MO), etc.), optical storage (such as CD, DVD, BD, HVD, etc.), and semiconductor memories (such as ROM, EPROM, EEPROM, non-volatile memory (NANDFLASH), solid state drive (SSD)), etc.

The device embodiments described above are only illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in One location, or it can be distributed across multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Persons of ordinary skill in the art can understand and implement the method without any creative effort.

Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the part of the above technical solution that essentially contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., including a number of instructions to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application, but not to limit it; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent substitutions are made to some of the technical features; however, these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions in the embodiments of the present application.

Claims (10)

1. An access request current limiting method, characterized by including: Obtain the label identification of the access request of the back-end service; the label identification is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request; Determine the target request whose number of requests for the access request exceeds the initial current limit value corresponding to the tag identification; Query the resource usage information of the first target server, and limit the flow of the target request according to the resource usage information; the resource usage information includes processor resource usage, disk usage, and memory usage. 2. The access request current limiting method according to claim 1, characterized in that the tag identification is also used to indicate the resource type required for the access request, and the resource type includes processor resources, memory resources and disks. Resources; limiting the target request according to the resource usage information includes: Determine the target resource type required by the target request according to the tag identification; Determine remaining resources of the first target server according to the resource usage information; Based on the target resource type and the remaining resources, determine whether the first target server has available resources corresponding to the target request; If it does not exist, obtain the target resource type and query whether there is a second target server in the back-end service; the second target server is the one that exists in the back-end service in addition to the first target server. The server of available resources corresponding to the target request; If there is a second target server in the backend service, the target request is flow-limited according to the available resources of the second target server. 3. The access request current limiting method according to claim 2, characterized in that the tag identifier is also used to indicate the timestamp information of the access request; the available resources of the second target server are Limit the flow of the above target requests, including: Obtain the target timestamp of the target request under the tag identification; Determine the window size based on the available resources of the second target server; Sort the target requests according to the order corresponding to the target timestamp; According to the sorting order of the target requests, a sliding window algorithm is used to respond to the target requests based on the window size to limit the flow of the target requests. 4. The access request current limiting method according to claim 2, characterized in that the second target server in the back-end service is queried according to the following formula whether there are available resources corresponding to the target request: R=αC+βM+δD; Among them, C is the processor usage, M is the memory usage, D is the disk usage, R is the resource abundance, and E _i is the i-th server in the back-end service except the first target server. The remaining resource indicator value, busness is the resource type preference factor corresponding to the target resource type, and fail _count is the number of failed requests within the historical preset time period; ε, α, β and δ are all preset weight parameters. 5. The access request current limiting method according to claim 1, characterized in that determining that the number of requests for the access request exceeds the target request of the initial current limiting value corresponding to the tag identification includes: Classify and summarize the access requests based on the label identification, and classify and summarize the access requests with the same label identification under the same label classification; According to the classification summary results, determine the number of access requests under the target label classification; the target label classification is any one of the label classifications corresponding to the label identification; If the number of requests exceeds the initial current limit value corresponding to the tag identifier under the target tag category, obtain the target requests under the target tag category that exceed the initial current limit value. 6. The access request current limiting method according to claim 1, characterized in that before obtaining the tag identification of the access request of the back-end service, it further includes: Obtain the request information of the access request of the back-end service; the request information includes the first target server accessed by the access request, the resource type required by the access request, and the timestamp information of the access request; Generate a label identification of the access request according to the request information. 7. The access request current limiting method according to claim 1, characterized in that the access request current limiting method further includes: Obtain resource usage information of historical access requests; The initial current limit value corresponding to the tag identifier of the historical access request is determined according to the resource usage information. 8. An access request current limiting device, characterized by including: A request interception module, used to obtain the tag identification of the access request of the back-end service; the tag identification is used to indicate the initial current limit value corresponding to the access request, and the first target server accessed by the access request; A first current limiting module, configured to determine target requests whose number of access requests exceeds the initial current limiting value corresponding to the tag identification; The second current limiting module is used to query the resource usage information of the first target server and limit the current flow of the target request according to the resource usage information; the resource usage information includes processor resource usage and disk usage. and memory usage. 9. An electronic device, comprising a processor and a memory storing a computer program, characterized in that when the processor executes the computer program, the access request current limiting method according to any one of claims 1 to 7 is implemented. step. 10. A non-transitory readable computer storage medium on which a computer program is stored. The characteristic is that when the computer program is executed by a processor, the access request current limiting method according to any one of claims 1 to 7 is implemented. A step of.